Fitting structural models to electricity prices

We develop a structural model for electricity spot prices that incorporates renewable generation into the multi-fuel bid-stack framework. We show that renewables can be modeled by adjusting the demand curve, effectively reducing the market to one with only conventional fuels. Under standard assumptions of inelastic demand and market equilibrium, we derive a general pricing model highlighting two special cases: a single conventional fuel and a proportional multi-fuel setting. Using German market data, we evaluate structural models within this framework and compare their out-of-sample performance with standard benchmark models, including machine-learning and ARX-type specifications. Our findings indicate that a theory-driven one-fuel structural model, using gas prices, EUA prices, and residual demand as driving variables, delivers the most robust out-of-sample performance. These results highlight the practical value of theoretical structural models, particularly their simplicity and interpretability.